Robust Control with Adaptive Law for the System Including Unobserved States

Model Reference Adaptive Control (MRAC) is consists of a reference model, a plant model, and adaptive law. It is made to follow the target system to the reference model by fixed gain. As study example, Yang developed an adaptive control for SISO state space system[1]. Yang’s method[1] was not discussed about the performance of the reference model. MRAC depends on performance of the reference model. It was important to consider the reference model. We adopted closed-loop system that was stabilized by Robust LQ Control for the reference model[2]. We extended Yang’s method to descriptor system, and verified by 2-DOF helicopter[2]. The system depends rationally on uncertain parameters, it is difficult to deal with the system. We adopted redundant descriptor representation to linear uncertainty. If a system has unobserved states, robust controller can’t be synthesized. Therefore, a reference model and an plant model are designed separately. Error between the actual plant and the reference model is canceled by adaptive law. The proposed method is applied to helicopter. For example problem, mass of helicopter is changed by luggage or people. Helicopter is oscillated by the weight. As a result, Crash accident is caused by this problem. It is deal with in the framework of MRAC. 3-DOF helicopter is used as experimental machine[3], [4], [5]. The effectiveness of our approach is verified by 3-DOF helicopter.